Google Git
Sign in
apache / tvm-vta / 9bef80c9b70aa389c7055dee1742601510f1d32c
commit9bef80c9b70aa389c7055dee1742601510f1d32c[log] [tgz]
authorLiangfu Chen <liangfu.chen@icloud.com>Mon Nov 12 11:24:48 2018 +0800
committerThierry Moreau <moreau@cs.washington.edu>Sun Nov 11 19:24:48 2018 -0800
treed1d98e812785679061e23b72dc5d648bdf16cf61
parent0fc5812a64b71419816e4d8595176e591fe2cb45 [diff]
[VTA] Improved RPC for VTA (#2043)

* assign default port to 9091 as the documented

* bug fix in printing RuntimeError and add an additional search path

* pretty print rebuild runtime args

* PRC => RPC

* replace vta_config.json file path

`build/vta_config.json` => `vta/config/vta_config.json`

* undo the change in adding lib_search path

* search vta_config.py file in vta/config

* avoid exposing driver function calls, and use predefined `VTAMemGetPhyAddr` instead.

* rename `tests/hardware/pynq` => `metal_test`

* set config path back to `build` dir
5 files changed
tree: d1d98e812785679061e23b72dc5d648bdf16cf61
  1. config/
  2. hardware/
  3. include/
  4. python/
  5. src/
  6. tests/
  7. tutorials/
  8. README.md
README.md

VTA: Open, Modular, Deep Learning Accelerator Stack

VTA (versatile tensor accelerator) is an open-source deep learning accelerator complemented with an end-to-end TVM-based compiler stack.

The key features of VTA include:

  • Generic, modular, open-source hardware
    • Streamlined workflow to deploy to FPGAs.
    • Simulator support to prototype compilation passes on regular workstations.
  • Driver and JIT runtime for both simulator and FPGA hardware back-end.
  • End-to-end TVM stack integration
    • Direct optimization and deployment of models from deep learning frameworks via TVM.
    • Customized and extensible TVM compiler back-end.
    • Flexible RPC support to ease deployment, and program FPGAs with the convenience of Python.

Learn more about VTA here.